bigPCAcpp (version 0.9.0)

pca_robust: Robust principal component analysis

Description

Compute principal component analysis (PCA) using robust measures of location and scale so that extreme observations have a reduced influence on the resulting components. The implementation centres each variable by its median and, when requested, scales by the median absolute deviation (MAD) before performing an iteratively reweighted singular value decomposition that down-weights observations with unusually large reconstruction errors.

Usage

pca_robust(x, center = TRUE, scale = FALSE, ncomp = NULL)

Value

A bigpca object mirroring the structure of pca_bigmatrix() with robust estimates of location, scale, and variance metrics.

Arguments

x

A numeric matrix, data frame, or an object coercible to a numeric matrix. Missing values are not supported.

center

Logical; should variables be centred by their median before applying PCA?

scale

Logical; when TRUE, variables are scaled by the MAD after centring. Scaling requires center = TRUE.

ncomp

Number of components to retain. Use NULL or a non-positive value to keep all components returned by the decomposition.

Examples

set.seed(42)
x <- matrix(rnorm(50), nrow = 10)
x[1, 1] <- 25  # outlier
robust <- pca_robust(x, ncomp = 2)
robust$sdev